The Agent Skills Directory

[SAFE]: The skill documentation and tool mappings align with the stated purpose of video editorial automation. No malicious behaviors, obfuscation, or unauthorized data access patterns were identified in the instruction set.
[INDIRECT_PROMPT_INJECTION]: The skill includes functionality for processing and searching transcribed spoken content from video clips, representing an attack surface for data-driven instructions.
Ingestion points: User-provided transcripts and spoken content are processed by the search_spoken_content and plan_transcript_tighten tools as described in SKILL.md.
Boundary markers: The skill explicitly implements a human-in-the-loop approach by recommending plan_dead_space_markers, which proposes reviewable markers for approval rather than executing silent destructive edits.
Capability inventory: The referenced tools perform timeline manipulation (duplicating, moving, and marking clips) within a running DaVinci Resolve instance via the davinci-resolve servers.
Sanitization: The skill requires the agent to provide explicit reasons for every proposed removal, facilitating manual audit and verification of the AI's interpretation of external data.

resolve-edit